Performance Evaluation And Prediction Of Typical Asphalt Pavement Structure In Common National And Provincial Trunk Highways

In recent years,as economy is developing rapidly,the traffic volume and the number of overloading and overweight vehicles in common national and provincial trunk highways increases rapidly,which cause severe performance degradation.Good pavement structure can improve the bearing capacity of the pavement structure,which effectively reduces the early damage of asphalt pavement and improves road comfort and durability.The quality of typical asphalt pavement structure in common national and provincial trunk highways is evaluated scientifically and objectively and the decay law of asphalt pavement performance is researched,which can provide reference for highway management departments to select pavement structure and formulate maintenance decisions.First of all,the pavement structures of the 9 common national trunk highways and 124 common provincial trunk highways in Guangdong province are analyzed statistically.The classification criteria of typical asphalt pavement structures in common national and provincial trunk highways in Guangdong province is determined based on the natural environmental and traffic load conditions.Typical asphalt pavement structures in common national and provincial trunk highways in Guangdong province are selected according to the characteristics of widely application and good prospects for development.Secondly,the performance evaluation index system of asphalt pavement is established and the performance of typical asphalt pavement structure in common national and provincial trunk highways in Guangdong province is evaluated by the standard evaluation method and the fuzzy entropy weight comprehensive evaluation method respectively.The ridge-type membership function is used for dealing with the fuzzy boundary of the pavement performance evaluation level,and the entropy method is introduced to solve the objectivity of empowerment in the fuzzy entropy weight comprehensive evaluation method.A reliable evaluation result about the performance is obtained by the fuzzy entropy weight comprehensive evaluation method which combines the advantages of the fuzzy method and the entropy method.Thirdly,the prediction models of pavement performance are established and the decay trend of pavement performance is studied by the time series analysis method,support vector machine algorithm and K nearest neighbor algorithm respectively.This paper explores the application of big data technology in pavement performance prediction.The time series analysis method has high precision in short-term prediction and under-fitting phenomenon in the longterm prediction.The support vector machine algorithm has obvious advantages in dealing with the linear indivisible problem in small samples,and can obtain the predicted value without the form and parameters of the prediction function.K-nearest neighbor algorithm has the characteristics of unique rapidity,convenience and accuracy in large-scale data classification and prediction.Finally,the performance prediction results of typical asphalt pavement structure based on support vector machine algorithm are used for evaluating the cost and benefit of typical asphalt pavement structure by cost-benefit ratio.Combined with the evaluation results of pavement performance and cost-benefit ratio,typical asphalt pavement structure in common national and provincial trunk highways of Guangdong Province is evaluated comprehensively.